Incidence, Burden on the Health Care System, and Factors Associated With Incisional Hernia After Trauma Laparotomy

Key Points

Question

What are the incidence, burden on the health care system, and factors associated with incisional hernia (IH) following trauma laparotomy (TL)?

Findings

This cohort study found a 10-year IH rate of 11.1% following TL that resulted in an additional 39.9% in aggregate costs relative to TL. Obesity, intestinal procedures, and repeated disruptions of the abdominal wall are among the strongest factors associated with IH.

Meaning

In this study, IHs were prevalent complications after TL with a similar patient profile to nontrauma populations; they resulted in considerable health care utilization and burdens to the health care system.

This population-based cohort study assesses the health care burden of and factors associated with incisional hernias after trauma laparotomy across health care settings.

Abstract

Importance

The evidence provided supports routine and systematic capture of long-term outcomes after trauma, lengthening the follow-up for patients at risk for incisional hernia (IH) after trauma laparotomy (TL), counseling on the risk of IH during the postdischarge period, and consideration of preventive strategies before future abdominal operations to lessen IH prevalence as well as the patient and health care burden.

Objective

To determine burden of and factors associated with IH formation following TL at a population-based level across health care settings.

Design, Setting, and Participants

This population-based cohort study included adult patients who were admitted with traumatic injuries and underwent laparotomy with follow-up of 2 or more years. The study used 18 statewide databases containing data collected from January 2006 through December 2016 and corresponding to 6 states in diverse regions of the US. Longitudinal outcomes were identified within the Statewide Inpatient, Ambulatory, and Emergency Department Databases. Patients admitted with International Classification of Diseases, Ninth Revision, Clinical Modification diagnosis codes for traumatic injuries with 1 or more concurrent open abdominal operations were included. Data analysis was conducted from March 2020 through June 2020.

Main Outcomes and Measures

The primary outcome was IH after TL. Risk-adjusted Cox regression allowed identification of patient-level, operative, and postoperative factors associated with IH.

Results

Of 35 666 patients undergoing TL, 3127 (8.8%) developed IH (median [interquartile range] follow-up, 5.6 [3.4-8.6] years). Patients had a median age of 49 (interquartile range, 31-67) years, and most were male (21 014 [58.9%]), White (21 584 [60.5%]), and admitted for nonpenetrating trauma (28 909 [81.1%]). The 10-year IH rate and annual incidence were 11.1% (95% CI, 10.7%-11.5%) and 15.6 (95% CI, 15.1-16.2) cases per 1000 people, respectively. Within risk-adjusted analyses, reoperation (adjusted hazard ratio [aHR], 1.28 [95% CI, 1.2-1.37]) and subsequent abdominal surgeries (aHR, 1.71 [95% CI, 1.56-1.88]), as well as obesity (aHR, 1.88 [95% CI, 1.69-2.10]), intestinal procedures (aHR, 1.47 [95% CI, 1.36-1.59]), and public insurance (aHRs: Medicare, 1.38 [95% CI, 1.20-1.57]; Medicaid, 1.35 [95% CI, 1.21-1.51]) were among the variables most strongly associated with IH. Every additional reoperation at the index admission and subsequently resulted in a 28% (95% CI, 20%-37%) and 71% (95% CI, 56%-88%) increased risk for IH, respectively. Repair of IH represented an additional $36.1 million in aggregate costs (39.9%) relative to all index TL admissions.

Conclusions and Relevance

Incisional hernia after TL mirrors the epidemiology and patient profile characteristics seen in the elective setting. We identified patient-level, perioperative, and novel postoperative factors associated with IH, with obesity, intestinal procedures, and repeated disruption of the abdominal wall among the factors most strongly associated with this outcome. These data support preemptive strategies at the time of reoperation to lessen IH incidence. Longer follow-up may be considered after TL for patients at high risk for IH.

Introduction

Incisional hernia (IH) is a prevalent and long-term complication^{1,2,3,4,5,6,7,8,9,10,11} with a considerable burden to any patient undergoing abdominal surgery, including those indicated for traumatic injuries. This chronic state¹² causes pain and detriment to quality of life and functional status,^13,14 and it carries a high recurrence and treatment-associated complications.¹⁵ Incisional hernia poses a considerable strain on health care systems, with 140 000 operations and a $7 billion expenditure for IH repair (IHR) annually.¹⁶ However, IH has been understudied, and important knowledge gaps exist when it comes to trauma, likely because of its emergency nature and the paucity of long-term postdischarge data,¹⁷ often exacerbated by noncompliance with follow-up¹⁸ in this population.

Despite the considerable impact of IH on trauma patients, only a few studies have explored the scope of IH after trauma laparotomy (TL). According to prior single-state, population-based studies,^19,20 IH rates range between 4% and 10.5%. However, these estimates relied solely on inpatient data, inevitably underestimating the true rate of IH, as evidenced by the rate of 13% to 25%^2,8,21 observed within nontrauma populations. In summary, several key questions remain unanswered. First, are IH incidence and patient profiles generalizable to the US population? Second, what is the burden on the health care system associated specifically with IH? Finally, what are the patient-level and operative factors associated with IH in patients undergoing TL? Current variables associated with IH stem from studies examining nontrauma populations^{22,23,24,25,26} and do not completely explore the complexity of trauma patients, who often require repeated disruptions of the abdominal wall through reoperations.

Therefore, the objective of this study was to determine the IH incidence after TL, its burden on the health care system, and factors associated with IH using data from different health care settings within geographically diverse regions in the US. We hypothesized that IH is prevalent among patients undergoing TL, there are patient-level and operative factors associated with IH, and the risk is higher among those who undergo reoperations during both index and subsequent hospitalizations.

Methods

Study Design and Population

This University of Pennsylvania institutional review board–approved, population-based cohort study used 18 statewide databases corresponding to 6 states. The data used were preexisting, publicly available, deidentified data, and therefore informed consent did not apply to this study. State Inpatient Databases (SIDs) from Florida (2006-2016; all date ranges are full calendar years), Iowa (2009-2016), Maryland (2013-2016), Nebraska (2006-2016), New York (2006-2016), and Wisconsin (2013-2016) served to identify patients admitted with traumatic injuries and concurrent open abdominal operations (defined by International Classification of Diseases, Ninth Revision, Clinical Modification [ICD-9-CM] diagnosis codes 800.xx-959.xx and procedure codes as per a prior study²⁶) without open abdominal procedures 6 months prior to the trauma encounter and/or active gastrointestinal cancer (ICD-9-CM codes 150.xx-159.xx) at the time of the trauma encounter. Those with trauma recidivism requiring laparotomy, a current IH diagnosis (either present on admission or acquired [ie, intentional IH after open abdomen]), IHR (eTable 1 in the Supplement) at the time of the index admission, an age younger than 18 years, and death at the index admission were excluded. Index encounters were identified until December 2014 to allow follow-up of 2 or more years.

Data Source and Covariates

The Healthcare Cost and Utilization Project (HCUP) SIDs were used to identify the index admission, primary outcome (IH occurrence), and readmissions. The State Ambulatory Surgery and Services Databases (SASDs) and State Emergency Department Databases (SEDDs) were used to extend the capture of the primary outcome to outpatient surgery, ambulatory services, observation stays, and emergency department (ED) encounters. For each state, the SIDs contain data on discharges for all patients admitted,²⁷ the SEDDs contain data on all ED discharges that did not result in an admission,²⁸ and the SASDs contain data on all ambulatory surgery and outpatient services (eg, observation stays, imaging).²⁹ These databases contain clinical and nonclinical variables on all patients regardless of age and insurance status and provide a unique visit link variable to allow indefinite follow-up across databases.

The variables extracted included age, sex, race, ethnicity, payer, income quartile, state, Injury Severity Score, and mechanism of injury (obtained using the International Classification of Diseases Program for Injury Categorization³⁰), Charlson Comorbidity Index score, smoking status, comorbidities (obesity, diabetes, hypertension, chronic lung disease, chronic heart failure, anemia, liver disease, kidney failure, coagulopathy, and weight loss, defined by HCUP-provided comorbidities or ICD-9 codes), length of stay, weekend admission, and disposition. Operative factors included procedure category or categories (esophagogastric, small bowel and/or large bowel, ostomy, hepatopancreaticobiliary, obstetrics/gynecology, urology, endocrine, or vascular), as defined by ICD-9 procedure codes described in a prior study.²⁶ Postoperative factors included surgical and nonsurgical complications (as described elsewhere³¹), a prolonged length of stay (>75th percentile), and the number of abdominal reoperations during the index admission and postdischarge.

Outcomes

The primary outcome was IH after TL diagnosed at the time of subsequent health care encounters. The cumulative incidence of IH was reported between 1 and 10 years after TL. An IH was defined by diagnosis and/or procedure codes (eTable 1 in the Supplement). The proportion of patients with IH captured at the time of an IHR encounter, emergency IHR, and IHR cases requiring mesh were ascertained. Secondary outcomes were overall and unplanned 1-year readmissions to any hospital within the state, readmission with bowel obstruction, hospital direct costs associated with inpatient IHR following TL, and number of IHRs. Only SID charges can be converted to costs using HCUP-provided cost to charge ratios. We conducted a subanalysis to estimate the patient-level and aggregate costs associated with inpatient IHRs. For patients with multiple IHRs, the aggregate readmission costs were obtained. These costs were normalized to 2019 dollars using the Consumer Price Index. We limited the sample until 2014, given the availability of charge-to-cost data at the time of the analyses.

Statistical Analysis

Descriptive statistics and annualized IH incidence rates per 1000 people were used to summarize characteristics. The overall cumulative incidence of IH was reported and plotted over time by the Kaplan-Meier method and number of reoperations. The cumulative incidence of IH by the number of reoperations was estimated through Cox regression, accounting for the time between the index and each subsequent abdominal procedure. Crude and risk-adjusted Cox proportional hazard regression models with robust standard errors clustering by hospital were used to determine factors associated with IH occurrence. The proportional-hazard assumption function was met. Covariates were selected based both on clinical judgment, factors associated with IH after abdominal surgery according to the literature, and significant baseline differences (ie, state, year). The potential collinearity between some HCUP-provided comorbidities and overlapping Charlson Comorbidity Index comorbidity scores was explored and found to be weak (R², 0.1-0.3) because their definitions differed. We accounted for the time-varying outcome of subsequent open abdominal procedures before IH or censoring. Florida and Maryland data sets did not provide discharge month, and thus, a discharge in June was assumed to allow censoring if the patient did not develop the outcome or died on readmission. The outcome of this assumption was assessed in a sensitivity analysis excluding patients from Maryland and Florida. Race and ethnicity were categorized as unreported for Nebraska, because these data were not provided. Missing data accounted for up to 4.33% for some of the variables of interest (insurance and income); a complete-case analysis approach was undertaken. Statistical significance set at P < .05. All analyses were performed using StataMP release 15 (StataCorp). Data analysis occurred from March 2020 through June 2020.

Results

After applying inclusion and exclusion criteria (eFigure in the Supplement), we identified 35 666 patients undergoing TL. Of these, 3127 (8.8%) developed an IH during the follow-up period (median [interquartile range (IQR)], 5.6 [3.4-8.6] years) and 2229 of 3127 (71.3%) were identified at the time of future IHR procedure encounters.

Overall, patients were a median (IQR) of 49 (31-67) years old, 21 014 (58.9%) were men, 21 584 (60.5%) were White, 49.8% had public insurance (Medicare, 11 349 [31.8%]; Medicaid, 6425 [18.0%]), and 28 909 (81.1%) were admitted for nonpenetrating trauma. The most prevalent comorbidity was hypertension (11 760 [33.0%]). The most common procedures were intestinal (12 974 [36.4%]) and hepatopancreaticobiliary (8338 [23.4%]). The overall and surgical complication rates were 61.0% (n = 21 760) and 46.7% (n = 16 647), respectively; 2129 patients (6.0%) required 1 abdominal reoperation and 647 patients (1.8%) more than 1. The remainder of characteristics assessed are shown in Table 1.

Table 1. Demographic, Clinical, and Operative Characteristics of Patients Admitted With Traumatic Injuries Who Underwent Emergency Laparotomy.

Characteristic	Patients, No. (%)
Overall	35 666
Age, median (IQR), y	49.0 (31.0-67.0)
Male	21 014 (58.9)
Female	14 652 (51.1)
Race and ethnicity
White	21 584 (60.5)
Black	6819 (19.1)
Hispanic	3585 (10.1)
Other	2094 (5.9)
Not reported	1584 (4.4)
Insurance
Medicare	11 349 (31.8)
Medicaid	6425 (18.0)
Private	12 069 (33.8)
Uninsured	3032 (8.5)
Other	2791 (7.8)
Income percentile
0-25	11 365 (31.9)
26-50	9022 (25.3)
51-75	8431 (23.6)
76-100	6848 (19.2)
State
Florida	18 656 (52.3)
Iowa	1061 (3.0)
Maryland	1520 (4.3)
Nebraska	1425 (4.0)
New York	12 068 (33.8)
Wisconsin	936 (2.6)
Injury Severity Score category
Mild trauma (0-7)	21 708 (60.9)
Moderate trauma (8-14)	6076 (17.0)
Severe trauma (>15)	7882 (22.1)
Mechanism of injury
Blunt	28 909 (81.1)
Penetrating injury	6757 (18.9)
Comorbidities
Obesity	2679 (7.5)
Diabetes	3594 (10.1)
Hypertension	11 760 (33.0)
Chronic pulmonary disease	5277 (14.8)
Chronic heart failure	2215 (6.2)
Deficiency anemia	6257 (17.5)
Liver disease	1020 (2.9)
Kidney failure	2213 (6.2)
Coagulopathy	3079 (8.6)
Weight loss	4953 (13.9)
Smoking	7107 (19.9)
Charlson Comorbidity Index score
0-2	32 027 (89.8)
>2	3639 (10.2)
Disposition
Home/routine	18 024 (50.5)
Transfer
Short-term hospital	895 (2.5)
Other long-term facilitya	10 474 (29.4)
Home with home health care	6021 (16.9)
Against medical advice	244 (0.7)
Weekend admission	8836 (24.8)
Procedure type
Esophagus and stomach	344 (1.0)
Small and/or large bowel	12 974 (36.4)
Hepatopancreaticobiliary	8338 (23.4)
Obstetrics and gynecology	3903 (10.9)
Urology	2223 (6.2)
Endocrine	2106 (5.9)
Vascular	2299 (6.4)
Ostomy	7054 (19.8)
Complication
Any	21 760 (61.0)
Surgical	16 647 (46.7)
Nonsurgical	12 507 (35.1)
Prolonged length of stay	8419 (23.6)
Reoperation
None	32 890 (92.2)
1	2129 (6.0)
>1	647 (1.8)

Abbreviation: IQR, interquartile range.

^aFor example, skilled nursing facilities and intermediate care facilities.

Outcomes of Multiple Abdominal Procedures

The yearly IH rates are shown in eTable 2 in the Supplement. These rose from 3.3% (95% CI, 3.1%-3.5%) at 1 year to 11.1% (95% CI, 10.7%-11.5%) at 10 years (Figure 1A). The median (IQR) time to IH was 1.4 (0.8-2.9) years. Of those who developed IH, more than one-third (1125 of 3127 [36.0%]) developed it by year 1, almost two-thirds by year 2 (1972 of 3127 [63.1%]) and 90.6% (n = 2832 of 3127) by year 5. Patients who underwent 0, 1, and more than 1 reoperations during the index admission demonstrated 10-year IH rates of 10.3% (95% CI, 9.9%-10.8%), 18.3% (95% CI, 16.5%-20.4%), and 27.3% (95% CI, 23.4%-31.6%), respectively (P < .001; Figure 1B). The same cumulative outcome was found at 10 years, considering subsequent abdominal reoperations after the index hospitalization with 10% (95% CI, 10%-10%), 22.3% (95% CI, 20.4%-24.5%), 34.2% (95% CI, 28.1%-41.3%) for those with 0, 1, and more than 1 subsequent reoperations, respectively (P < .001; Figure 1C). Examination of reoperations as a continuous variable demonstrated that the risk of IH increased by 28% (95% CI, 20%-37%) for each additional reoperation at the index admission and 71% (95% CI, 56%-88%) for each subsequent abdominal surgery postdischarge.

Figure 1. Cumulative Incidence of Incisional Hernia (IH) Following Emergency Trauma Laparotomy.

A, Overall. B, By number of abdominal reoperations during index admission. C, By number of subsequent abdominal reoperations after discharge.

Factors Associated With IH

Table 2 shows the annualized incidence rates and crude hazard ratios (HRs) by patient and operative characteristic factors. The overall IH annual incidence rate was 15.6 (95% CI, 15.1-16.2) per 1000 people, and there were factors significantly associated with increased IH. All postoperative factors assessed resulted in significantly higher incidence or risk for IH (Table 3). Reoperations at the index hospitalization resulted in the greatest annual incidence rates for IH, reaching 83.4 per 1000 people (95% CI, 66.6-104.5 per 1000 people; HR, 5.16 [95% CI, 3.81-6.99]; P < .001). The association of factors with IH according to Cox regression modeling are shown in Figure 2. Notably, obesity (adjusted HR [aHR], 1.88 [95% CI, 1.69-2.10]), the number of subsequent abdominal procedures (aHR, 1.71 [95% CI, 1.56-1.88]), intestinal procedures (aHR, 1.47 [95% CI, 1.36-1.59]), liver disease (aHR, 1.47 [95% CI, 1.21-1.78]), public insurance status (aHRs: Medicare, 1.38 [95% CI, 1.20-1.57]; Medicaid, 1.35 [95% CI, 1.21-1.51]), and reoperations at the index admission (aHR, 1.28 [95% CI, 1.2-1.37]) were associated with increased risk of IH, whereas the Charlson Comorbidity Index score (aHR, 0.91 [95% CI, 0.88-0.94]), female sex (aHR, 0.91 [95% CI, 0.84-0.98]), vascular procedure (aHR, 0.82 [95% CI, 0.71-0.95]), and Black race (aHR, 0.58 [95% CI, 0.52-0.66]) were associated with decreased IH rates. These findings persisted within a sensitivity analysis.

Table 2. Incisional Hernia Cases and Annual Rates per 1000 People and Crude Hazard Ratio by Demographic, Clinical, and Operative Characteristics in Patients Admitted for Trauma Who Underwent Emergency Laparotomy.

Characteristic	Total patients, No. (%)	Annual rate of incisional hernia per 1000 individuals (95% CI)	Hazard ratio (95% CI)	P value
Overall	35 666 (8.8)	15.6 (15.1-16.2)	NA	NA
Age range, y
18-45	15 835 (7.0)	11.5 (10.9-12.2)	1 [Reference]	NA
46-65	10 043 (12.5)	23.5 (22.2-24.8)	1.95 (1.79-2.13)	<.001
>65	9788 (7.7)	15.2 (14.2-16.3)	1.28 (1.16-1.43)	<.001
Sex
Male	21 014 (9.3)	16.5 (15.8-17.2)	1 [Reference]	NA
Female	14 652 (8.1)	14.4 (13.6-15.2)	0.87 (0.81-0.93)	<.001
Race and ethnicity
White	21 584 (9.9)	18.0 (17.2-18.8)	1 [Reference]	NA
Black	6819 (6.0)	10.5 (9.6-11.6)	0.58 (0.52-0.65)	<.001
Hispanic	3585 (9.2)	16.2 (14.5-18.0)	0.89 (0.80-1.00)	.045
Other	2094 (6.4)	10.7 (9.0-12.7)	0.60 (0.51-0.71)	<.001
Not reported	1584 (6.9)	12.2 (10.1-14.7)	0.60 (0.44-0.82)	.002
Insurance
Medicare	11 349 (8.3)	16.3 (15.3-17.4)	1 [Reference]	NA
Medicaid	6425 (8.0)	14.3 (13.1-15.6)	0.87 (0.77-0.97)	.01
Private	12 069 (9.8)	16.6 (15.7-17.6)	1.05 (0.96-1.15)	.30
Uninsured	3032 (7.4)	11.7 (10.3-13.4)	0.78 (0.65-0.93)	.005
Other	2791 (9.8)	16.3 (14.5-18.3)	1.06 (0.88-1.29)	.54
Income percentile
0-25	11 365 (15.2)	14.5 (13.6-15.4)	1 [Reference]	NA
26-50	9022 (0.7)	15.3 (14.3-16.5)	1.07 (0.96-1.19)	.22
51-75	8431 (1.6)	16.7 (15.6-17.9)	1.15 (1.04-1.28)	.007
76-100	6848 (1.6)	16.6 (15.4-18.0)	1.16 (1.04-1.30)	.008
Injury Severity Score category
Mild trauma (0-7)	21 708 (8.5)	15.3 (14.6-16.0)	1 [Reference]	NA
Moderate trauma (8-14)	6076 (8.4)	14.6 (13.4-15.9)	0.96 (0.87-1.07)	.49
Severe trauma (>15)	7882 (9.7)	17.3 (16.1-18.6)	1.12 (1.01-1.24)	.03
Mechanism of injury
Blunt	28 909 (9.1)	16.5 (15.9-17.1)	1 [Reference]	NA
Penetrating injury	6757 (7.5)	12.3 (11.3-13.5)	0.76 (0.67-0.86)	<.001
Comorbidities
Obesity	2679 (14.9)	32.6 (29.6-36.0)	2.02 (1.81-2.25)	<.001
Diabetes	3594 (9.3)	18.5 (16.6-20.6)	1.16 (1.04-1.31)	.01
Hypertension	11 760 (10.2)	20.0 (18.9-21.2)	1.39 (1.28-1.51)	<.001
Chronic pulmonary disease	5277 (10.2)	20.3 (18.7-22.1)	1.32 (1.21-1.44)	<.001
Chronic heart failure	2215 (7.1)	14.8 (12.6-17.3)	0.93 (0.80-1.09)	.36
Deficiency anemia	6257 (8.9)	17.0 (15.6-18.4)	1.08 (0.98-1.19)	.12
Liver disease	1020 (12.5)	26.3 (22.1-31.3)	1.58 (1.32-1.90)	<.001
Kidney failure	2213 (8.0)	17.9 (15.5-20.8)	1.11 (0.96-1.29)	.17
Coagulopathy	3079 (9.8)	20.3 (18.1-22.7)	1.24 (1.09-1.42)	.001
Weight loss	4953 (9.8)	21.3 (19.5-23.3)	1.32 (1.17-1.48)	<.001
Smoking	7107 (10.0)	18.8 (17.5-20.2)	1.22 (1.11-1.34)	<.001
Charlson Comorbidity Index score
0-2	32 027 (8.9)	15.6 (15.0-16.1)	1 [Reference]	NA
>2	3639 (7.4)	16.5 (14.7-18.6)	0.97 (0.85-1.11)	.66
Disposition
Home/routine	18 024 (7.5)	12.6 (11.9-13.3)	1 [Reference]	NA
Transfer
Short-term hospital	895 (9.9)	19.0 (15.4-23.4)	1.53 (1.23-1.89)	<.001
Long-term facilitya	10 474 (8.5)	16.8 (15.7-17.9)	1.30 (1.17-1.45)	<.001
Home with home health care	6021 (12.9)	23.6 (22.0-25.4)	1.84 (1.67-2.03)	<.001
Against medical advice	244 (5.7)	9.2 (5.4-15.5)	0.76 (0.45-1.30)	.32
Weekend admission	8836 (8.2)	14.5 (13.5-15.6)	0.90 (0.83-0.98)	.02
Procedure
Esophagus and stomach	344 (8.1)	14.8 (10.2-21.4)	0.94 (0.63-1.41)	.78
Small or large bowel	12 974 (11.8)	22.0 (20.9-23.1)	1.77 (1.65-1.91)	<.001
Hepatopancreaticobiliary	8338 (9.8)	16.9 (15.7-18.1)	1.13 (1.04-1.22)	.004
Obstetrics and gynecology	3903 (4.3)	6.9 (5.9-8.0)	0.42 (0.36-0.50)	<.001
Urology	2223 (8.0)	14.4 (12.5-16.7)	0.91 (0.76-1.08)	.28
Endocrine	2106 (1.9)	3.1 (2.3-4.2)	0.18 (0.12-0.26)	<.001
Vascular	2299 (7.3)	12.9 (11.1-15.0)	0.83 (0.71-0.96)	.02
Ostomy	7054 (10.5)	20.6 (19.1-22.1)	1.40 (1.27-1.54)	<.001

Abbreviation: NA, not applicable.

^aFor example, skilled nursing facilities and intermediate care facilities.

Table 3. Incisional Hernia Cases and Annualized Rates per 1000 People Among Patients Undergoing Emergency Laparotomy for Trauma, by Postoperative Factor.

Postoperative factors	Total patients, No. (%)	Annual rate of incisional hernia per 1000 individuals (95% CI)	Hazard ratio (95% CI)	P value
All patients	35 666 (8.8)	15.6 (15.1-16.2)	NA	NA
Any complication
No	13 906 (6.8)	11.3 (10.6-12.0)	1 [Reference]	NA
Yes	21 760 (10.1)	18.7 (18.0-19.5)	1.63 (1.50-1.78)	<.001
Surgical complicationa
No	19 019 (7.4)	12.8 (12.1-13.5)	1 [Reference]	NA
Yes	16 647 (10.3)	19.2 (18.3-20.1)	1.48 (1.37-1.60)	<.001
Nonsurgical complication
No	23 159 (8.0)	13.6 (13.0-14.3)	1 [Reference]	NA
Yes	12 507 (10.2)	19.8 (18.8-21.0)	1.43 (1.31-1.55)	<.001
Prolonged length of stay
No	27 247 (7.8)	13.6 (13.0-14.1)	1 [Reference]	NA
Yes	8419 (11.8)	23.2 (21.8-24.7)	1.68 (1.54-1.84)	<.001
Nonroutine discharge
No	18 024 (7.5)	12.6 (11.9-13.3)	1 [Reference]	NA
Yes	17 642 (10.0)	19.2 (18.3-20.1)	1.50 (1.37-1.63)	<.001
Reoperation
None	32 890 (8.1)	14.2 (13.7-14.8)	1 [Reference]	NA
1	2540 (15.6)	31.6 (28.6-34.8)	2.15 (1.94-2.37)	<.001
≥2	236 (32.2)	83.4 (66.6-104.5)	5.16 (3.81-6.99)	<.001

Abbreviation: NA, not applicable.

^aThe surgical complication composite includes surgical site infection, hematoma, seroma, wound dehiscence, fistula, delayed healing, bleeding complicating procedure, unintentional laceration or puncture, gastrointestinal complications (ileus, obstruction, gastric dilation, persistent vomiting, and noninfectious gastroenteritis), reoperation, blood transfusion, and other, nonspecified postoperative complications. Nonsurgical complications included cardiovascular, respiratory, genitourinary, and deep venous thromboembolism complications. The 1-year readmission outcome was measured after the index admission.

Figure 2. Forest Plot Showing the Independent Association of Patient-Level, Operative, and Postoperative Factors With Incisional Hernia (IH) Occurrence Following Emergency Trauma Laparotomy.

In addition to the factors listed above, the model was controlled for state and year and accounted for clustering at the hospital level. Unadjusted (blue) and adjusted (orange) point estimates (squares) are shown with their 95% CIs (line). HR indicates hazard ratio; OB/GYN, obstetrics and gynecology.

Health Care Utilization Associated With IHR

Of the 35 666 patients, 2229 (6.2%) underwent IHR. Among this cohort, 1895 (85.0%) underwent only 1 repair, 273 (12.2%) underwent 2 repairs, and 61 (2.7%) had 3 or more repairs, totaling 2644 IHR operations. Of these operations, 1710 (64.7%) involved placing mesh, and 404 (15.3%) occurred as emergencies. Subanalyses of inpatient costs of IHR included 1320 patients, of whom 1264 (95.7%) had cost data available. In this sample, the median (IQR) costs for inpatient IHR admissions were $16 629 ($10 186-$31 137) and totaled $36.1 million. These costs represented an increase of 39.9% of the aggregate costs of the TL admissions included in the cost analysis ($90.6 million).

Readmissions and Bowel Obstruction

The 1-year overall and unplanned readmission rates were 67.9% (24 205 patients of 35 666) and 37.9% (13 508 of 35 666), respectively. The 1-year, 5-year, and 10-year rates of bowel obstruction were 7.8% (95% CI, 7.5%-8.1%), 13.7% (95% CI, 13.4%-14.1%), and 13.1% (95% CI, 16.5%-17.5%), respectively.

Discussion

Using data corresponding to 35 666 patients who were admitted in 6 states in diverse US regions; underwent TL; and were followed up for up to 11 years within inpatient, outpatient, and ED encounters, we found a median time to IH of 1.4 years, and most IHs occurred within 5 years (90.6%), accumulating to a 10-year rate of 11.1%. After discharge, IH represented a substantial source of health care utilization, since a large proportion of subsequent IH cases were identified at the time of their surgical repair (71.3%), 1 in 6 (15.3%) of IHR operations occurred as emergencies, and IHR procedures resulted in a 40% increment to longitudinal inpatient health care costs. Moreover, we identified and quantified what are, to our knowledge, new operative and postoperative factors that were also associated with IH and demonstrated the cumulative outcome of abdominal reoperations. This study provides a population-based perspective that has important clinical and epidemiological implications for the chronic health state that represents IH.

Incisional hernia is a frequent, yet understudied,¹⁹ clinically significant complication after abdominal surgery. Its incidence is reported to range from 4% to 39% after elective procedures^{1,2,3,4,5,6,7,8,9,10,11} and 4% to 19% in the trauma setting.^19,20,32,33 In this study, the 10-year rate of IH was 11.1%. The variation in reported IH rates is highly dependent on the population assessed, follow-up time, data, and approach. Our results are in line with prior studies that reported IH rates of 10.5% in a cohort at 6.7 years¹⁹ and 6.3% at 5 years.³² Bowie et al²⁰ reported a lower rate of 4%, but this large discrepancy may be because of the use of inpatient data from a single US state, the exclusion of reoperations, and a different follow-up time. Instead, we combined data from several states, and this novel (to our knowledge) and comprehensive method captured IH not only when patients returned for IHR in the inpatient or outpatient setting but also when they presented for unrelated reasons for ambulatory services, ED visits, or readmissions. This approach captured 29% of IH cases that would have not been captured otherwise. Nonetheless, with approximately 1 in 9 patients developing IH after TL, half of patients being publicly insured, 1 of every 6 IHRs occurring in the emergency setting, and the 39.9% addition to all TL index admission costs that resulted when patients require IHR, IH represents a public health problem that merits attention and additional efforts to alleviate its burden on the health care system.

Beyond quantifying the IH incidence, we illustrate the time dependence of IH at a large scale. These findings have important policy and research implications. More than one-third of IHs occurred within 1 year and almost two-thirds within 2 years, and its incidence plateaus after 91% occur within 5 years, which is line with prior studies.^2,34 These findings support the recommendation of 5 years as the optimal follow-up time for studying IH,³⁵ with a minimum follow-up of 2 years. Furthermore, we corroborated the morbidity of IH seen in nontrauma settings extends to the trauma population independent of their younger age³⁶ and lower comorbidity burden.

This study also provides useful information on the factors associated with IH and confirms characteristics in the patient profile seen in other populations. Recognizing that most TL cases occur as emergencies and shortly after patients’ arrival renders preoperative characteristics associated with IH less germane. Nonetheless, there are also perioperative factors (ie, type of surgery) and postoperative factors (ie, complications) associated with IH. Factors associated with IH have been identified in prior studies,^{22,23,24,25,26} and some of them are associated with impaired wound healing (ie, smoking); however, only a single study examined them specifically for trauma among US military members.³³ In agreement, intestinal procedures were independently associated with IH. Our study extended beyond the operative phase and found that postoperative complications and reoperations increased the risk for IH. Indeed, repeated abdominal wall integrity disruption through additional procedures both during the index admission and subsequent encounters resulted in a substantial increase in the risk of IH formation among survivors of TL. Each reoperation during the index admission increased the risk of IH by 28% (95% CI, 20%-37%). Similarly, every additional subsequent abdominal surgery increased the risk of 71% (95% CI, 56%-88%). While other studies have excluded these complex patients,²⁰ they were purposefully included both to resemble what occurs in real-world management of their cases and illustrate the cumulative outcomes of reoperations on long-term IH formation. This is the first time, to our knowledge, that longitudinal outcomes of subsequent abdominal surgeries have been investigated in trauma surgery at a population-based level. These findings confirm data from prior smaller studies^2,37 and underscore the need for reconsidering surveillance and risk reduction strategies for these patients particularly in the postdischarge period. Based on these findings, surgeons may lengthen follow-up time for these patients and consider counseling on increased risk for IH during postoperative visits, which may have a positive association with compliance and encourage patients to return if they develop an IH. Furthermore, surgeons may leverage existing tools for risk reduction and optimization of underlying health conditions³⁸ and perhaps use enhanced closure techniques at the time of surgery (ie, prophylactic mesh placement³⁹) for prevention of IH at or before future elective abdominal procedures in patients at high risk, in addition to evidence-based closure strategies that reduce the incidence of IH, such as the small-bites technique⁴⁰ or barded suture.⁴¹ Future studies would do well to examine the association of intraoperative factors in addition to prophylactic measures with reducing IH development.

Abdominal wall reconstruction is an important part of the recovery for patients who develop IH after TL. In this cohort, almost 70% of the patients underwent IHR. Since the development of IH can be an indolent health process, it is not surprising to see that most patients were captured at the time of their repair. While the operative details used during IHR were beyond the scope of this study (ie, type of incisional hernia repair, type of mesh, reconstructive plane), we did discover that mesh was used in 69.5% of them. This highlighted the complexity of abdominal wall reconstruction after TL.

Limitations

The limitations of this study stem from its retrospective approach and the use of administrative data, which lack granular clinical data and may be subject to coding bias. For instance, we were unable to discern through billing codes TL cases left with open abdomen intentionally during the index admission. This bias was minimized by excluding those with active IH diagnosis or IHR at the time of the index admission. Also, these data do not include office visits and therefore exclude patients who may have been diagnosed with IH in clinic, particularly for asymptomatic hernias, loss of follow-up, or patients who relocated out of state. This bias resulted in underestimating the IH incidence and may overestimate the proportion necessitating surgical repair. Nonetheless, by linking the SID, SEDD, and SASD data, we comprehensively included inpatient admissions, ED visits, observation stays, and outpatient encounters at any hospital within the state in which an IH diagnosis was made or where an IHR procedure took place. This approach had never been used before to examine the IH epidemiology in the US, to our knowledge. Lastly, it is unknown whether the findings observed in this study apply to the entire US population. However, the results should closely resemble national outcomes, because this large cohort included patients from diverse geographical US regions (Northeast, South, and Midwest) across multiple health care settings with no fewer than 2 years and up to 11 years of follow-up.

Conclusions

We have demonstrated that the time dependence, patient profile, and health care burden of IH after TL mirrors those of the nontrauma setting. Incisional hernias develop long after surgery at a 10-year rate of 11.1%, with most cases occurring by 5 years. Incisional hernia repair following TL increases aggregate costs by approximately 40% relative to all index admissions. There are patient-level, operative, and postoperative factors associated with IH; obesity, intestinal procedures, and reoperations both at the index admission and postdischarge are among the strongest ones. The evidence supports routine and systematic capture of long-term outcomes after trauma, lengthening the follow-up for patients at risk for IH after TL, counseling on the risk of IH during the postdischarge period, and consideration of preventive strategies before future abdominal operations to lessen IH prevalence, as well as the patient and health care burden.

Supplement.

eTable 1. Definitions of incisional hernia and incisional hernia repair by ICD-9 and CPT codes.

eTable 2. Incisional hernia (IH) cases and annual rates per 1,000 people and crude hazard ratio by demographic, clinical and operative characteristics in patients admitted for trauma undergoing emergent laparotomy.

eFigure. Flow diagram of inclusion and exclusion criteria.